Efficient Approach to Segment Ligatures and Open Characters in offline Arabic Text

Abstract

Abstract-This paper researches offline Arabic handwriting recognition. It introduces a new approach to segmentation ligature and open Arabic character based on the structural perspective dealing with sub-words/words, including dots to recognize individual letters. Segmentation approaches that have been integrated into the recognition phase have the capability to deal with ligatures and closed characters issues. This complex problem is due to the cursive writing nature of the Arabic language. This paper also develops an Arabic character algorithm on segmented, pixel based and centre reign (CR) to recognize the letters. The evaluation results are stated in the IFN/ENIT and IAM database which indicate the recognition rate and the effectiveness of our system.

Fast Technique to Recognize Closed Characters in Arabic Handwriting

“…This section is devoted to shedding light on numerous articles that introduced many methods to improve the efficiency of recognizing Arabic characters. Such methods include LHT [6], [7], [8], [ 9], [ 10] , HMM [11], [12], [13], [14], histogram with Gabor filter [15], histogram with projection profile [16], [17], Otsu's model [9], [17], [18], heuristic rules differentiating [19], discrete cosine transform [20], parallel thinning algorithm [21], artificial immune system [22], and/or dynamic Bayesian network [23]. In addition, morphological algorithms [18] were applied to estimate document skew angles using the IFN/ENIT database.…”

“…Reference [16] presented algorithms that solved the problems of ligature, overlapping, and open characters. In addition, they divided the method of segmentation into open-and semi-open loop characters.…”

A deep learning based approach for extracting Arabic handwriting: applied calligraphy and old cursive

“…The distance among the start and finish pixels in the contouring segmentation phase may impact the x -axis, y -axis, and structural feature characteristics in contextual text extraction. The training and evaluation materials and associated graphics also describe structural elements such as left, right, top, and bottom directions ( Saber et al, 2017 ). Several works ( Obaidullah et al, 2017 ; Roy et al, 2016 ; Dhall et al, 2011 ; Rodríguez & Perronnin, 2008 ; Chherawala, Roy & Cheriet, 2013 ; Terasawa & Tanaka, 2009 ; Lu, Liong & Zhou, 2017 ; Ayyalasomayajula, Nettelblad & Brun, 2016 ; Sun et al, 2016 ; Yuan & Liberman, 2008 ) investigate the extraction of characteristics including dot–number connections in individual segments, trains, branching, and secondary strokes (height-to-width and inclines from the initial to the final phase) ( Roy et al, 2016 ).…”

“…The baseline, as a key, serves as a delimiter for horizontal projections in addition to script localization, including the essential functionality that detects the direction of the beginning point. To differentiate modifiers in zones among the well-known approaches, optical character recognition (OCR) ( Srihari, Shekhawat & Lam, 2003 ) is regarded as an acknowledged (as well as the standard) method for grasping and analyzing many languages with varying features and difficulties, such as Chinese ( Du & Huo, 2013 ; Patil & Shimpi, 2011 ), English ( Bag, Harit & Bhowmick, 2014 ; Saber et al, 2017 ). Text recognition systems with high recognition rates have been created by researchers.…”

“…For the Arabic recognition task, it is your responsibility to extract Arabic text from photos. These varying styles of writing were, on the other hand, retained as intricate patterns and materialized as visuals or films ( Saber et al, 2017 ). Handwriting recognition technologies are often used to decipher written text from a picture.…”